Jeff Kell ([log in to unmask]) wrote:
: Well, it's nothing "drastic" unless you get a bit more exotic, but you should
: be aware of these things if you're mixing and matching legacy MPE with Posix:
 
1.
  : shell/iX> ccposix one two three
  : TZ=[GMT0], TERM=[hp2392a]            [set by HP-supplied /etc/profile ]
 
2.
  : shell/iX> CCMPE one two three            [Odd parm value, but ctime is
  : TZ=[], TERM=[]                            Note info string here, no info
 
3.
  : (MANAGER.SYS): ./ccposix "one two three"   [Now ctime is right (without
  : TZ=[], TERM=[]
 
4.
  : (MANAGER.SYS): CCMPE "one two three"
  : TZ=[], TERM=[]
 
The difference for the variables is explained for each case:
 
  1.  Program used /lib/libc.a, getenv uses posix environment.
      TZ and TERM are initialized by the shell (see /etc/profile).
 
  2.  Program used LIBC.LIB.SYS, getenv uses CI variables.
      TZ and TERM are not defined in your CI variables.
 
  3.  Program used /lib/libc.a, getenv uses posix environment.
      Posix environment inherited from CI variables.
      TZ and TERM are not defined in your CI variables.
 
  4.  Program used LIBC.LIB.SYS, getenv uses MPE CI variables.
      TZ and TERM are not defined in your CI variables.
 
Try the test again after
 
  :setvar TZ "GMT0"
  :setvar TERM "hp2392a"
 
and you will find that they are all the same again.
 
BTW, the program name thing is basically a bug that wasn't caught.  It returns
a null for HFS names because the underlying FFILEINFO call returns an empty
name when you ask for the MPE name and it isn't representable as such.
 
Steve